Mapping of Planetary Surface Age Based on Crater Statistics Obtained by an Automatic Detection Algorithm
Total Page:16
File Type:pdf, Size:1020Kb
The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XLI-B4, 2016 XXIII ISPRS Congress, 12–19 July 2016, Prague, Czech Republic MAPPING OF PLANETARY SURFACE AGE BASED ON CRATER STATISTICS OBTAINED BY AN AUTOMATIC DETECTION ALGORITHM Atheer L. Saliha, M. Mühlbauera, A. Grumpea, J. H. Pasckertb, C. Wöhlera, H. Hiesingerb aImage Analysis Group, TU Dortmund, 44227 Dortmund, Germany – {atheer.altameemi | maximilian.muehlbauer | arne.grumpe | christian.woehler}@tu-dortmund.de bInstitut für Planetologie, Westfälische Wilhelms-Universität Münster, 48149 Münster, Germany – {jhpasckert | hiesinger}@uni-muenster.de Commission IV, WG IV/8 KEY WORDS: Crater statistics; CSFD; automatic crater detection; absolute model age; age mapping ABSTRACT: The analysis of the impact crater size-frequency distribution (CSFD) is a well-established approach to the determination of the age of planetary surfaces. Classically, estimation of the CSFD is achieved by manual crater counting and size determination in spacecraft images, which, however, becomes very time-consuming for large surface areas and/or high image resolution. With increasing availability of high-resolution (nearly) global image mosaics of planetary surfaces, a variety of automated methods for the detection of craters based on image data and/or topographic data have been developed. In this contribution a template-based crater detection algorithm is used which analyses image data acquired under known illumination conditions. Its results are used to establish the CSFD for the examined area, which is then used to estimate the absolute model age of the surface. The detection threshold of the automatic crater detection algorithm is calibrated based on a region with available manually determined CSFD such that the age inferred from the manual crater counts corresponds to the age inferred from the automatic crater detection results. With this detection threshold, the automatic crater detection algorithm can be applied to a much larger surface region around the calibration area. The proposed age estimation method is demonstrated for a Kaguya Terrain Camera image mosaic of 7.4 m per pixel resolution of the floor region of the lunar crater Tsiolkovsky, which consists of dark and flat mare basalt and has an area of nearly 10,000 km2. The region used for calibration, for which manual crater counts are available, has an area of 100 km2. In order to obtain a spatially resolved age map, CSFDs and surface ages are computed for overlapping quadratic regions of about 4.4 x 4.4 km² size offset by a step width of 74 m. Our constructed surface age map of the floor of Tsiolkovsky shows age values of typically 3.2-3.3 Ga, while for small regions lower (down to 2.9 Ga) and higher (up to 3.6 Ga) age values can be observed. It is known that CSFD-derived absolute model ages can exhibit variations although the surface has a constant age. However, for four 10-20 km sized regions in the eastern part of the crater floor our map shows age values differing by several hundred Ma from the typical age of the crater floor, where the same regions are also discernible in Clementine UV/VIS color ratio image data probably due to compositional variations, such that the age differences of these four regions may be real. 1. INTRODUCTION In this paper an automatic crater detection algorithm (CDA) is Impact crater counting is one of the most important tools for applied to the mare-like floor region of the lunar farside crater estimating the geologic age of a planetary surface, i.e. the Tsiolkovsky, in order to obtain crater counts of similar accuracy period of time that has passed since the last resurfacing event. as manual crater counts. A surface age map is created based on Recent planetary spacecraft missions have provided enormous the automatically obtained crater counts, resulting in a spatially amounts of high-resolution image data of global coverage for resolved map of the AMA of the examined surface area. various solar system bodies, allowing for the construction of crater catalogues comprising large numbers of small craters 2. CRATER DETECTION ALGORITHMS (Stepinski et al., 2012; Robbins, 2009, 2016). In the domain of planetary science, crater counts are typically Impacts of small bodies on planetary surfaces lead to the performed by manual inspection of orbital images (e.g., Michael formation of craters (Hörz et al., 1991). Impact craters are and Neukum, 2010). On the other hand, an important research especially abundant on Mercury, the Moon and Mars (Grieve et topic in remote sensing is the development of crater detection al., 2007), where they persist over long periods of time due to algorithms (CDAs) which automatically determine the locations the lack of plate tectonics, an atmosphere, and life. For and sizes of craters in images (Salamanićcar and Lončarić, planetary bodies without an atmosphere, the areal impact crater 2008, 2010). Numerous CDAs have been developed as a density increases with increasing surface age, i.e. densely consequence of the high importance of knowledge about the cratered surface parts are usually older than surface parts with a impact crater distribution for remote geologic studies of low abundance of craters. The impact crater size-frequency planetary surfaces (Stepinski et al., 2012). However, many distribution (CSFD), denoting the diameter-dependent areal works about CDAs are limited to the demonstration of the density of impact craters for a given surface part, allows for an accuracy of a particular algorithm on a small set of test images estimation of the so-called absolute model age (AMA) of the displaying quite simple and clearly pronounced craters, while surface (Hartmann, 1999; Hartmann and Neukum, 2001; meaningful studies in the field of planetary science require Michael and Neukum, 2010; Michael et al., 2012; Michael, CDAs achieving a high performance also for less obvious, e.g., 2013, 2014, 2015; Hiesinger et al., 2011). degraded, impact craters (Stepinski et al., 2012). This contribution has been peer-reviewed. doi:10.5194/isprsarchives-XLI-B4-479-2016 479 The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XLI-B4, 2016 XXIII ISPRS Congress, 12–19 July 2016, Prague, Czech Republic Since the start of the last decade, there was a further increase in The template matching stage is followed by a fusion process research works regarding the automatic detection of craters that replaces multiple detections of the same crater at slightly (Salamunićcar and Lončarić, 2010). Existing CDAs can be different positions or with slightly different diameters by the divided into two different groups. CDAs of the first group average of these positions or diameters. detect craters in images, while CDAs of the second group detect craters based on digital elevation models (Salamunićcar and Lončarić, 2010; Di et al., 2014). 4. THE REGION OF INTEREST: TSIOLKOVSKY According to Stepinski et al. (2012), some image-based CDAs The crater Tsiolkovsky (Figure 3) is one of a small number of consist of techniques which process the image to detect the lava-filled lunar farside craters, thus having a very dark and characteristic circular or elliptical shapes of crater rims, smooth crater floor. It is located at (20° S, 129° E) and has a followed by a matching stage such as the Hough transform diameter of approximately 200 km. The central peak located in (Hart, 2009), while other CDAs employ machine learning the northern part of the mare-filled region has a “W” shape approaches for crater detection. These systems are trained using (Tyrie, 1988a). image data sets labelled by human experts, and the detector performance is tested on new image data not used during the Previous studies regarding Tsiolkovsky crater suggest ages training stage (Stepinski et al., 2012, see also Chung et al., around 3.8 Ga (Walker and El-Baz, 1982) and 3.5±0.1 Ga 2014). Examples for classifier architectures are boosting (Tyrie, 1988b) for the floor region. The image used in this work techniques allowing for a combination of various feature was a mosaic of images taken by the Terrain Camera (TC) of detectors (Wetzler et al., 2005; Salamunićcar and Lončarić, the lunar orbiter spacecraft Kaguya (Haruyama et al., 2014) 2010; Di et al., 2014; Stepinski et al., 2012) and support vector with a resolution of 7.4 m/pixel. The dark mare fill inside machines (Wetzler et al., 2005; Chung et al., 2014). Tsiolkovsky crater is our area of interest, as it appears to be geologically homogeneous. In this work, the template matching based CDA introduced by Grumpe and Wöhler (2013) with the ability of automatic The estimation of the absolute model age (AMA) mainly detection of small craters (<10 image pixels diameter) in lunar depends on the number of craters per diameter interval and per images will be used for crater detection. This template-based area, the CSFD. We chose a surface part of around 100 km2 size algorithm will be applied to orbital image data of the mare-like inside the crater Tsiolkovsky as a reference dataset for the crater floor region of lunar crater Tsiolkovsky. The main step in the detection, which has been evaluated manually by Pasckert et al. estimation of the absolute surface age is the statistical analysis (2015). The crater diameter interval considered for estimating of the CSFD constructed using the image-based CDA for the AMA ranges from 128 m to 1000 m. Smaller craters were determining the locations and diameter values for the derivation not taken into account because the detection rate of the CDA of the AMA. Accordingly, a spatially resolved AMA map of the decreases with decreasing diameter for diameters less than whole floor of Tsiolkovsky will be constructed. about 100 m, leading to an artificial “roll-off” of the CSFD.